The effectiveness of dimensionless scaling strategies was assessed using temporal-spatial data collected from an anthropometrically diverse group of participants over a range of walking speeds. Video analysis of children (aged 4-15 years, mean = 10 years) and adults (18-40 years, mean = 25.2 years), each walking at their freely chosen speed, showed adults to take significantly longer strides than children at any given speed (predominately due to their longer lower limbs). Regression analysis of stride length versus walking speed showed that the slopes for adults and children were similar, but that the intercept was significantly higher in adults. Childrens' data were more scattered compared to those for adults. Plots of relative stride length (L') versus dimensionless speed (u') reduced intra-group variation and eliminated significant differences between adults and children, although subtle differences occurred between children of different ages. These findings support the use of dimensionless scaling in gait analysis, but care should be taken when using dimensionless numbers in relation to children under about 10 years of age due to ineffectiveness of scaling strategies in this group. Normalization for differences in stature using dimensionless scaling was also effective for participants walking at speeds significantly above or below their freely chosen (= 'most efficient') walking speed, suggesting a broad applicability for assessing participants who are unable to walk at their normal walking speed (e.g., participants with endoprostheses, osteoarthritis, or various musculoskeletal problems).